Formulation and validation of a predictive model to correct blood glucose concentrations obtained with a veterinary point-of-care glucometer in hemodiluted and hemoconcentrated canine blood samples.

Abstract

To determine the effect of PCV on veterinary point-of-care (POC) glucometer measurements in canine blood samples and develop a formula to correct the glucose concentration as measured by a point-of-care glucometer (POCgluc) given a known PCV. Experimental and prospective study. Blood samples from 6 healthy dogs and from 30 hospitalized dogs. 60 mL of heparinized blood was obtained from each of 6 healthy dogs. Samples were processed into packed RBCs and plasma. Packed RBCs were resuspended with plasma to achieve a range of PCVs from 0% to 94%. Duplicate POCgluc and PCV measurements were obtained for each dilution; following POCgluc measurements, plasma samples were analyzed for glucose concentration by a clinical laboratory biochemical analyzer (LABgluc). A correction formula for POCgluc was developed. Measurements of POCgluc, PCV, and LABgluc were also determined from blood samples of 30 dogs admitted to the veterinary teaching hospital. Values of LABgluc for each sample were similar at any PCV. As PCV decreased, POCgluc was falsely increased; as PCV increased, POCgluc was falsely decreased, compared with LABgluc. The absolute difference between POCgluc and LABgluc increased as the PCV changed from 50%. Compared with POCgluc, the corrected POCgluc had a significantly improved correlation with LABgluc, which was also reflected in improvements in Clarke and consensus error grid analyses. Results indicated that in dogs with hemodilution or hemoconcentration, POCgluc did not reflect actual patient glucose concentrations. Use of a correction formula reduced this error. Corrected POCgluc data had strong, significant correlations with LABgluc data.